Structure extraction from texture via relative total variation

TLDR

Meaningful structures form on textured surfaces, but extracting them is challenging due to varied texture patterns, yet it is practically important. The study proposes new inherent variation and relative total variation measures and an efficient optimization system to extract main structures. The authors validate the new variation measures on millions of sample patches and employ an efficient optimization system to extract structures. The approach enables manipulation, rendering, and reuse of many structure‑with‑texture images and drawings that were previously difficult to edit.

Abstract

It is ubiquitous that meaningful structures are formed by or appear over textured surfaces. Extracting them under the complication of texture patterns, which could be regular, near-regular, or irregular, is very challenging, but of great practical importance. We propose new inherent variation and relative total variation measures, which capture the essential difference of these two types of visual forms, and develop an efficient optimization system to extract main structures. The new variation measures are validated on millions of sample patches. Our approach finds a number of new applications to manipulate, render, and reuse the immense number of "structure with texture" images and drawings that were traditionally difficult to be edited properly.